Project Halo

The consortium and the project goals

Seattle-based Vulcan Inc (http://www.vulcan.com) has created the long-term research program Project Halo (http://www.projecthalo.com) to develop a “Digital Aristotle” that can serve as a comprehensive computer knowledge base and problem-solving system for the natural sciences. Project Halo is developing technologies for low-cost, highly-scalable modeling of basic scientific knowledge -- primarily physics, chemistry, and biology at the American AP level -- and developing methods for automatically answering sophisticated scientific questions posed in English. Applications of Project Halo technology include automated tutors for students, and knowledge-based "associate" systems for supporting scientific work.

Vulcan has selected ontoprise GmbH (headquartered in Karlsruhe, Germany) as its prime contractor to develop an extension to the Semantic MediaWiki for Project Halo; the extension is named halo extension and is part of ontoprise's commercial semantic enterprise wiki distribution named SMW⁺ (ref. SMW+). SMW⁺ is serving as a web-based collaboration tool for the user-friendly and low-cost acquisition of scientific knowledge for Project Halo.

The ontoprise team consists of experts from ontoprise itself and its project partners at the Institute AIFB of the University of Karlsruhe, and the Saarbrücken-based German Research Center for Artificial Intelligence.

Figure 1 shows to which extend project halo helped improving Semantic MediaWiki in terms of interoperability and usability compared to competing semantic wikis.

Figure 1: project halo improved usability and interoperability of SMW to outpace competing semantic wikis.

Quick links to project resources

Please consult the following resources for more information:

• halo Extension
  • Overview of the features of the halo Extension: halo Extension
  • A demonstration video of the extension's main features is available at http://www.smwplus.net/index.php/Tutorial/SMW%2B_Demo_video.
  • User community of Semantic MediaWiki and the halo Extension: http://smwplus.net
  • Download the halo Extension from sourceforge: sourcefourge.
  • Enter bug reports here [1] (please note: you need an account in the smwforum first: [2])
  • Read about installation, features and latest release in the SMW user forum for commercial and production use: [3].
• SMW⁺ easy-to-install distribution of Semantic MediaWiki (and further extensions)
  • Overview of the features of SMW⁺: SMW+.
  • SMW⁺ home page describing features and benefits: [4]
  • Get your copy of SMW⁺ here: [5].

The "halo extension" - a extension to make Semantic MediaWiki user friendly

Vulcan tasked Ontoprise and its partners to increase the reliability and performance of Semantic MediaWiki, as well as create the open-source extension halo extension that supports the following goals:

Browse and edit the wiki's ontology in the web-based OntologyBrowser.

Dramatically increase the size of the Semantic MediaWiki author community by lowering the entry barrier and increasing the usability of the core system. This included building: Syntax highlighting to easily access the wiki text visually; WYSIWYG editor to allow untrained users to author wiki articles; A semantic tool bar allows easier annotation of the wiki text; Autocompletion for all properties, categories and templates; and Improved navigation in the ontology with the GUI-based ontology browser.

Embed semantic tags into the rendered wiki article.

Supporting the authoring of of more and deeper knowledge specifically tuned to the natural sciences. This included building: Support of chemical formulas and -equations and annotation with SI-units of measurement, New types of relations such as inverse, transitive and symmetric, Support of cardinalities and domain and range annotations for relations and Increased expressivity through n-ary relations (included in SMW)

Lucene-based retrieval interface combining full-text hits and semantic data (SKOS meta data).

Building in new techniques to allow users to detect inconsistencies and continuously improve the quality of the authored knowledge. This included building: Extensive search functions with basic ask query reasoning; and Many new "semantic gardening" features and background checkers

Graphical query interface to build ask-queries graphically.

Improving the ability to access the knowledge in Semantic MediaWiki. This included building: An external mapping tool to facilitate reuse of the wiki knowledge base; and A GUI-based question formulation interface to easily assemble inline queries.

Graphical tool to connect external web services to wiki articles.

Improving the connectivity and importing features in Semantic MediaWiki. This included building: a web service connector with web-based GUI to call external web services from the wiki and integrate the web service results directly into your wiki articles. a MS-Excel add-in to query an knowledge base and import the results into Excel spreadsheets (uses SPARQL) a term import framework enables the flexible import of structured data and vocabularies into your wiki. a triple store connector (for Jena and OntoBroker) to augment inline queries with inverse, transitive and symmetric relations.

All parts of the extension are subject to formal usability evaluations, which resulted in several improvements to the UI and workflows. Please consult Halo Extension for a detailed description of the software extension including how to install it. An in-depth description of all features of the halo extension can be found in Features of the Halo Extension.

The context of Semantic MediaWiki in Project Halo

The configuration of tools and methods within Project Halo includes three main components:

1. A semantic wiki for broad-based knowledge modeling (at the RDF(S)+ level of expressivity);
2. A Rich Modeling Environment, which allows for deep modeling of scientific knowledge which cannot be expressed in the wiki semantics (such as process flow knowledge), question answering, problem solving, and explanation generation.
3. A powerful mapping tool which allows the knowledge in the Semantic MediaWiki to be exploited by the Rich Modeling Environment.

The basic use scenario includes overlapping knowledge formulation and question formulation phases. The shared semantic wiki is used during knowledge formulation by all wiki-users to author, store, critique, and share knowledge. The generated knowledge (KB) from the wiki can be exported as RDF and reused with other software products. Furthermore the KB can be integrated into a Rich Modeling Environment (RME) which allows advanced reasoning with the wiki data – such as Project Halo's AURA or Ontoprise’s standard product OntoStudio®. This is done by mapping the wiki ontology to the internal ontology of the RME with the mapping tool. This mapping process enriches the knowledge captured in the RME with additional wiki-based concepts, instances, relations or attributes. This additional knowledge can be accessed by users by browsing the knowledge base or by querying.

Project schedule

The part of Project Halo follows this schedule:

• Stage 1
  • Project Kickoff: December 15th, 2006
  • Design Complete: April 1st, 2007
  • Feature Freeze: August 1st, 2007
  • Delivery: October 15th, 2007
• Stage 2
  • Project Kickoff: November 1st, 2007
  • Delivery: February 15th, 2008
• Stage 3
  • Project Kickoff: March 1st, 2008
  • Design Complete: May 1st, 2008
  • Feature Freeze: July 15th, 2008
  • Delivery: September 1st, 2008
• Stage 4
  • Project Kickoff: October 1st, 2008
  • Design Complete: December 23rd, 2008
  • Feature Freeze and deliveries: in Q1, Q2, Q3 and Q4 2009

About Vulcan

Vulcan Inc. creates and advances a variety of world-class endeavors and high-impact initiatives that change and improve the way people live, learn, do business and experience the world. Founded in 1986 by investor and philanthropist Paul G. Allen, and under the direction of president and CEO Jody Patton, Vulcan oversees various business and charitable projects including real estate holdings, investments in more than 40 companies, including Charter Communications, DreamWorks Animation SKG, Digeo Broadband, the Seattle Seahawks NFL and Portland Trail Blazers NBA franchises, First & Goal Inc., Vulcan Productions, the Seattle Cinerama theatre, Experience Music Project, the Science Fiction Museum & Hall of Fame, the Allen Institute for Brain Science and the Paul G. Allen Family Foundation. For more information about Vulcan, please visit http://www.vulcan.com.

About Ontoprise

Ontoprise is a leading software provider of ontology-based solutions. Ontologies structure relevant Know-how of your company, no matter how complex the relations are and how widely the information is scattered throughout the enterprise. The multi-patented technology enable information integration across heterogeneous sources, thus enterprise data will become a commonly available knowledge source for each and every employee with the appropriate need-to-know level. For more information about Ontoprise, please visit http://www.ontoprise.de.